Summary The Air Canada Jazz Bombardier CRJ705 aircraft (registration C-FBJZ, serial number15037) was operating as JZA8105 from Houston, Texas, to Calgary, Alberta, with 64passengers and 4crew members on board. The aircraft entered Canadian airspace in the vicinity of COUTS intersection, while in cruise at flight level400. Shortly thereafter, at 2046mountain daylight time, the crew contacted Edmonton Centre and was advised that traffic was being metered into Calgary. The controller gave the crew a crossing time for the VUCAN intersection of 2121and offered the option of slowing at pilot's discretion. This crossing time was 18minutes later than planned. The crew was concerned about fuel reserves as the flight had already been diverted west of its planned route in the United States due to weather and air traffic control requirements. The pilot reduced power to slow the aircraft in response to the anticipated metering of traffic. While the crew was occupied with the fuel calculations for a possible hold, the airspeed deteriorated to the point of stick shaker activation. The aircraft speed was recovered by applying power and initiating a descent. There was no damage to the aircraft or injuries to passengers or crew. The flight landed at Calgary without further incident. Ce rapport est galement disponible en franais. Other Factual Information At the time of the occurrence, the aircraft's weight of about 68000pounds and its centre of gravity were within approved limits. There were no known defects that might have influenced the aircraft's performance. The aircraft was clear of cloud at the time of the occurrence, and there was little or no turbulence. Forecast upper winds at flight level (FL)400 were 260True(T) at 18knots and temperature -55C. The incident occurred during the inaugural week of the company's scheduled CRJ705operation. Both pilots had previous experience in the CRJ200model (CL-600-2B19), and had completed a conversion course for the CRJ705 (CL-600-2D15). The conversion course consisted of ground instruction, simulator training, and a pilot proficiency check (PPC). The captain completed initial CRJ200series training in September2004 and the CRJ705conversion training in March2005. His last recurrent CRJ200PPC was in February2005. The first officer received his CRJ200 type endorsement in October2004, and his CRJ705 conversion in May2005. His last recurrent PPCwas in February2005. Both crew members arrived in Houston earlier the same day on JZA8106 at approximately 0200mountain daylight time (MDT).1 The captain flew the aircraft with a different pilot acting as first officer, with the occurrence first officer deadheading on the aircraft. The crew arrived at their hotel at 0300. The captain slept fitfully for 3.5hours before rising at 0830and was not well rested before the flight to Calgary. Take-off for the return flight (JZA8105) to Calgary was at1725. Fatigue impairs alertness and performance, and fatigued persons may fix their focus on minor problems, fail to anticipate danger, display flawed logic, or apply inappropriate corrective actions.2 Fatigue can significantly impair performance on safety-critical tasks. Included among the common effects of fatigue are the following: reduced attention; preoccupation with single tasks or elements of a task; and overlooked or misplaced sequential task elements. In anticipation of traffic metering into Calgary, the captain, who was the pilot flying, reduced power to slow from the cruise speed of 0.77M3 to the long range cruise speed of0.70M. The flight director was set to altitude track mode with the autopilot engaged. The captain then focused attention to the flight management system (FMS) on the centre console to help the first officer determine fuel reserves for a possible hold. About 90seconds later, the first officer noticed the speed dropping below 0.70Mand warned the captain. The captain responded by advancing the thrust levers. The aircraft continued to decelerate with increasing nose-up pitch attitude while maintaining FL400. The captain then began programming a descent in the flight control computer (FCC) vertical speed mode. Immediately thereafter, as the airspeed decreased to 0.61M(180knots indicated airspeed (KIAS)), the stick shaker activated and the autopilot disengaged. Both pilots pushed the control yoke forward to reduce the pitch attitude, which resulted in a descent and an increase in airspeed. This was followed by the crew returning the aircraft to a pitch-up attitude, with an increase in body angle of attack and G.4 A second stick shaker activation occurred 11seconds after the first. Buffeting and roll oscillations of about 10 degrees accompanied the stick shaker events. The pitch attitude was further reduced and the airspeed recovered to 0.82M(250KIAS). The altitude stabilized briefly at FL386 before the crew coordinated withATC, a further descent to FL380 due to conflicting traffic. (See AppendicesAandB for digital flight data recorder plots.) Drag versus speed (adapted from Air Canada Jazz training module) Minimum drag speed (Vmd) is defined as the speed at which aircraft total drag is at a minimum, requiring minimum thrust for level flight. This speed is associated with the endurance speed. At a constant altitude, progressive reduction in true airspeed below Vmd results in a concomitant increase in total drag, which results in a further deterioration of airspeed. Therefore, airspeed values below Vmd can result in negative stability (see Figure1). Should cruise airspeed become less than Vmd, speed restoration can be accomplished by increasing engine power, by initiating a descent, or a combination of both. The power required to maintain Vmd increases as altitude and/or aircraft weight increases. At high altitudes, the power required to maintain Vmd may be only marginally lower than that required for cruise. Delaying a descent in favour of applying power may result in the aircraft speed decelerating to the point of stick shaker activation and disengagement of the autopilot. Data supplied by Bombardier Aerospace indicated that, under the conditions JZA8105 was operating at the time of the occurrence, Vmd was 0.75M, which was near the cruise speed of 0.77M. In the CRJ200, Vmd was below the published long range cruise speed under all normal flight conditions, whereas in the CRJ705, under certain altitudes and aircraft weights, Vmd could be above the long range cruise speed. An Operational Evaluation Board Report (OEBR) prepared by Transport Canada in 2002, and revised in2004, established master differences in training, pilot checking, and currency requirements for pilots operating CL-65 series aircraft. Operator aircraft differences requirement (ODR) tables in the OEBR contained detailed model/variant differences and training requirements. Canadian commercial operators, including Air Canada Jazz, could use the report to develop flight crew training programs and checks in accordance with the Canadian Aviation Regulations (CARs). Flight characteristics relating to aircraft aerodynamic responses to airspeed changes were included in definitions. However, because the flight characteristics were considered to be similar, the ODR tables did not include differences between the CRJ200and CRJ705/900 related to Vmd. The Air Canada Jazz pilot training syllabus included sections devoted to high altitude physiology and depressurization in accordance with CARs. During the initial type training on the CRJ200, company pilots received training in high-altitude/high-speed aerodynamics specific to the type. Approach to stall recognition and recovery training was limited to 10000feet above sea level (asl) and did not cover the significant differences in techniques between low-altitude and high-altitude speed recovery. In June2005, simulator training in airspeed recovery at high altitude, which emphasized timely pitch reduction, was added to the CRJ recurrent training syllabus. Both occurrence pilots had completed their recurrent training before that time and therefore had not received this training. Before the incident, the company's CRJ705 Aircraft Operating Manual included a recommended normal holding speed of 225KIAS for altitudes above 14000feet. There was no reference to the fact that this speed could be substantially below Vmd at altitudes near the maximum cruising altitudes. Although data related to Vmd was included in the Bombardier Flight Planning and Cruise Control Manual, the information was not readily available to flight crews in the cockpit. It was not included in quick reference material or aircraft operating manuals. Due to the length of the flight remaining, the cockpit voice recorder information pertaining to the incident had been overwritten. The digital flight data recorder (DFDR) was removed and analysed by the TSB Engineering Laboratory. It was determined that, during the occurrence flight, the offset between the recorded DFDR time and the GPS-based Coordinated Universal Time (UTC) changed by one second about every 10minutes. At the time of the incident, the time recorded by the DFDR was 23seconds longer than the recorded UTC time. This corresponded to a difference between timing parameters of 0.16 per cent. Section625.33 of the CARs stipulates that the accuracy limits for recorded time drift be within 0.125 per cent per hour. The observed drift suggested that the stability requirements as set out by the CARs were exceeded. Since the GPS-based UTC time stamp would have been very accurate, it is likely that the time drift originated in the data concentrator unit (DCU) (part number DCU-4004) of the DFDR. This information was communicated to the aircraft manufacturer and the component supplier for evaluation.